fix: 修复 RAG 无限循环问题和导入错误
Some checks failed
构建并部署 AI Agent 服务 / deploy (push) Has been cancelled

主要修复:
1. 修复 RAG 推理无限循环问题(大小写不匹配 + 缺少已检索结果检查)
2. 修复 intent_classifier.py 的绝对导入错误
3. 删除旧的 start.sh 脚本,添加新的启动脚本
4. 优化路由逻辑和状态管理
This commit is contained in:
2026-05-04 18:59:15 +08:00
parent 9841f47432
commit c9bf21be0e
13 changed files with 503 additions and 164 deletions

View File

@@ -8,10 +8,10 @@ from typing import Dict, Any, Optional, List
from dataclasses import dataclass, field
from datetime import datetime
from app.main_graph.state import MainGraphState
from app.logger import info, debug
from app.model_services.chat_services import get_small_llm_service, get_chat_service
from app.main_graph.nodes.rag_nodes import rag_retrieve_node
from ..state import MainGraphState
from ...logger import info, debug
from ...model_services.chat_services import get_small_llm_service, get_chat_service
from .rag_nodes import rag_retrieve_node
# ========== 核心数据类型 ==========
@@ -367,8 +367,8 @@ async def fast_rag_node(state: MainGraphState, config: Optional[Dict[str, Any]]
debug(f"[Fast RAG] 发送事件失败: {e}")
try:
# 先尝试 RAG 检索
state = rag_retrieve_node(state, config)
# 先尝试 RAG 检索 - 注意rag_retrieve_node 是异步函数,需要 await
state = await rag_retrieve_node(state, config)
# 检查检索结果
rag_docs = getattr(state, "rag_docs", [])